neurovlm.train.which_device

neurovlm.train.which_device()

Determine the device to move models and tensors to.

Returns:

device

Return type:

{“cuda”, “mps”, “cpu”}